or Connect
AppleInsider › Forums › Mac Hardware › Future Apple Hardware › The Intel Powermac / Powermac Conroe / Mac Pro thread
New Posts  All Forums:Forum Nav:

The Intel Powermac / Powermac Conroe / Mac Pro thread - Page 2

post #41 of 947
Quote:
Originally posted by hmurchison
I'm seeing a continuing and disturbing trend here that defies logic.

Conroe AND Woodcrest share the same core. They both have the same 14 pipeline stage Core Architecture. The primary difference is that Woodcrest supports SMP and has a 1333 FSB to Conroe's 1066, support Dual Independent Bus and perhaps twice the L2(which I am beginning to doubt that it will at launch)

Thus Conroe in not really a consumer only processor. In fact Yonah is more likely the new consumer desktop Conroe the midrange and Woodcrest the highend.

No you won't see all Woodcrest Powermac replacements because to benefit you'd have to have dual socket motherboards and that is expensive.

The iMac may not take conroe that well because its TDP is 65 watts. That's pretty hot for the iMacs thin chassis. We'll see what Apple engineers can do.

If you're running a computer with a single Conroe vs a single Woodcrest the differences in speed will be very minute. The two have the same core so any other differences would come down to cache, FSB and perhaps effect from the DIB. Very small in the real world just as a XEON isn't that much faster than a Pentium 4 until you move to SMP systems.

To address the part that I highlighted in bold. I thought that would be the Memrom processor. I could be wrong.


Back on PowerMac track:

What I figured was once Apple does it's usual scaling back on the new systems Conroe would go into a PowerMac in the low end model as the one single socket model. My logic behind this is because that has been the trend with the PowerMac for the past few years. Apple has been releasing all Dual socket systems at the start to satisfy the major range of their highend market, and then after sales have gotten to a point they throw in an update for the low end model with a single socket to let another portion of their buying market get their hands on a tower if they so wish. I don't see why they would stop doing that. Especially now being that the Conroe is going to come out after the Woodcrest is anyway. This coincidentally turns out like perfect timing for Apple's highend market selling strategy.

On to iMac again:

Another thing to remember is that these processors can come in various speed ranges. If they can separate the models by getting a lower Hz frequency out of a conroe that wont burn up an iMac I'd have to guess they would use Yonah Core Duo's in the iBook, and Mac mini, Memrom's in a MacBook Pro, and a Tablet (unless it's the same product with a swivel screen (((I HOPE))), and a Memrom in the low level 15" iMac, and a scaled back and cooler Conroe for the bigger iMacs. Putting a conroe into an iMac seems unlikely, but I guess it could be a possibility.
Although you can't forget we are dealing with intel now, and not Moto, or IBM so new processors will be rolling out with different speeds, and specs in great abundance. By the time Apple were to get around to feeling they need a Conroe to put life back into iMac sales intel should have a newer processor perfect for the job that we all will be anticipating.
onlooker
Registered User

Join Date: Dec 2001
Location: parts unknown




http://www.apple.com/feedback/macpro.html
Reply
onlooker
Registered User

Join Date: Dec 2001
Location: parts unknown




http://www.apple.com/feedback/macpro.html
Reply
post #42 of 947
and why not 3 models of "PowerMacs":
- under $2000 with a single Conroe chip 1066 FSB
- under $3000 with a single Woodcrest chip 1333 FSB
- under $4000 with two Woodcrest chips 1333 FSB

with 2 options for each model:
- $1499 2.40GHz Conroe ($316)
- $1999 2.67GHz Conroe ($530) + better GPU/biggerHD/more RAM

- $2299 2.67GHz Woodcrest ($700) + better GPU
- $2499 3.00GHz Woodcrest ($850)

- $2999 dual 2.67GHz Woodcrest ($1400)
- $3299 dual 3.00GHz Woodcrest ($1700)

(cpu list price)

Even before Woodcrest, Apple could build a quad Mac with the Yonah-based Sossaman (Xeon LV, 667FSB, 2M cache) chips:
- quad 1.66GHz Xeon LV (2x$209) --- $1999
- quad 2.00GHz Xeon LV (2x$423) --- $2499
It's not a 64-bit system, but there are some applications that are not 64-bit but are "core/processor" aware that would greatly benefit from quad-core (even @ 32-bit) systems...

Just to say that in the months to come Apple will (finally) have a lot of chips to choose from to build Macs.
Will they be bold enough to please everybody with lower cost headless Macs?
We'll certainly know at the WWDC.
post #43 of 947
Quote:
Originally posted by mjteix
and why not 3 models of "PowerMacs":
- under $2000 with a single Conroe chip 1066 FSB
- under $3000 with a single Woodcrest chip 1333 FSB
- under $4000 with two Woodcrest chips 1333 FSB

with 2 options for each model:
- $1499 2.40GHz Conroe ($316)
- $1999 2.67GHz Conroe ($530) + better GPU/biggerHD/more RAM

- $2299 2.67GHz Woodcrest ($700) + better GPU
- $2499 3.00GHz Woodcrest ($850)

- $2999 dual 2.67GHz Woodcrest ($1400)
- $3299 dual 3.00GHz Woodcrest ($1700)

(cpu list price)

Even before Woodcrest, Apple could build a quad Mac with the Yonah-based Sossaman (Xeon LV, 667FSB, 2M cache) chips:
- quad 1.66GHz Xeon LV (2x$209) --- $1999
- quad 2.00GHz Xeon LV (2x$423) --- $2499
It's not a 64-bit system, but there are some applications that are not 64-bit but are "core/processor" aware that would greatly benefit from quad-core (even @ 32-bit) systems...

Just to say that in the months to come Apple will (finally) have a lot of chips to choose from to build Macs.
Will they be bold enough to please everybody with lower cost headless Macs?
We'll certainly know at the WWDC.

I doubt if we'll see Sossman.

But, I would be more than willing to pay $4,000 for a dual Woodcrest machine. Esp. if it had either Crossfire, or SLI, though I prefer ATI boards. While Nvidia is slightly better for games, ATI is better for 3D work. Either is more than good enough for 2D, though, again, ATI is much better for video.

Also, I'd love to see 2 removable front bays again. No excuse for machines at this price level, and size. The G5's had room for one more half height drive above the DVD. It could have been a replaceable bay.

You know, my first Mac was a 950. It cost, in early 1992, $6,000. That didn't include the keyboard ($189), or a CD player because there wasn't any for internal mounting. I bought a Toshiba 2x speed, which was the fastest model at the time ($999), and reworked the computer in my machine shop downstairs. I was the first person in the country (at least) to have an internal CD-ROM.

The point is that many are willing to pay more to get more.

Jobs has screwed that up since he first came out with the Mac, and he's been screwing it up ever since he came back. The pro machines have been getting less and less versatile. He only put two optical bays back on the later G4's because of all the screaming, and took them away as soon as he could. It's a shame. There is NO reason why a $2,000 to $3,300 machine shouldn't have 6 slots and four internal drive bays, at least 2 of which would be external.

I remember when a Powermac came out with 3 slots, and he was asked why he didn't release a bigger machine like the older ones, the 9600, for example (which I also have).

His response that that would be "only" 5% of their customers. We, in the business were agog! Apple was selling almost 4 million machines a year, at the time (before they fell further). 5% would have been about 200 thousand, at the then price of over $5,500. That would have been $1.1 billion !

Foolish man.

I hope he's come to his senses now.
post #44 of 947
There usually are 3 models of powerMac's but this one here
Quote:
- under $2000 with a single Conroe chip 1066 FSB
- under $3000 with a single Woodcrest chip 1333 FSB

Is pretty much the same machine.

If you go to the Apple store there is usually 3, or 4 configurable tiers to chose from for the PowerMac. But I'm pretty sure they will add a conroe version either once the processor is ready, or after they reach a certain point in sales. But a single woodcrest makes little sense. Conroe is almost the same processor as the woodcrest, but it's designed for single socket use. There really is no reason to hamper the abilities of a woodcrest by using it in a single socket with conroe coming right after it.

BTW Apple tried the headless Mac, and it was their biggest failed computer in years. Even with the elegant design of the cube there was little demand for it, and it was thusly discontinued almost overnight.

Quote:
Originally posted by melgross
While Nvidia is slightly better for games, ATI is better for 3D work. Either is more than good enough for 2D, though, again, ATI is much better for video.


I'm not sure what gave you that idea, but the Quadro is by far the best 3D graphics card for pro 3D use.
onlooker
Registered User

Join Date: Dec 2001
Location: parts unknown




http://www.apple.com/feedback/macpro.html
Reply
onlooker
Registered User

Join Date: Dec 2001
Location: parts unknown




http://www.apple.com/feedback/macpro.html
Reply
post #45 of 947
Quote:
Originally posted by onlooker
There usually are 3 models of powerMac's but this one here
Is pretty much the same machine.

If you go to the Apple store there is usually 3, or 4 configurable tiers to chose from for the PowerMac. But I'm pretty sure they will add a conroe version either once the processor is ready, or after they reach a certain point in sales. But a single woodcrest makes little sense. Conroe is almost the same processor as the woodcrest, but it's designed for single socket use. There really is no reason to hamper the abilities of a woodcrest by using it in a single socket with conroe coming right after it.

BTW Apple tried the headless Mac, and it was their biggest failed computer in years. Even with the elegant design of the cube there was little demand for it, and it was thusly discontinued almost overnight.



I'm not sure what gave you that idea, but the Quadro is by far the best 3D graphics card for pro 3D use.

The Cube had poor marketing an Apple trademark.

Apple failed to make sure that people knew that it was not only fully upgradeable, but that it had an extra slot as well. People looked at it, and didn't understand that. When I would explain it to them, they would be very surprised.
post #46 of 947
Thread Starter 
I hope that the Powermacs will be smilarly priced to the current models, and I ESPECIALLY hope that they don't try bullshit graphics card markups.
post #47 of 947
Quote:
Originally posted by onlooker
I'm not sure what gave you that idea, but the Quadro is by far the best 3D graphics card for pro 3D use. [/B]

There would be plenty of disagreements on that one, believe me. But, it's the ONLY high end board available for us, unfortunately.
post #48 of 947
Quote:
Originally posted by Placebo
I hope that the Powermacs will be smilarly priced to the current models, and I ESPECIALLY hope that they don't try bullshit graphics card markups.

That's actually my hope. But, if we got more of a machine, I would pay more.

I would even buy a 4 chip machine, if they came out with one. Jobs wants to use Macs in Pixar. He's said that. It would have to be pretty good for the render farm.
post #49 of 947
Quote:
Originally posted by melgross
The Cube had poor marketing an Apple trademark.

Above all, they priced the Cube very poorly. It was more expensive and less expandable than the equivilent PowerMac G4s of the time. They thought people would pay a significant premium for the compact size and near-silent operation, but they were wrong.

The Mac mini shows they learned their lessons from the G4 Cube: it's aimed at clear market segments that don't overlap other products and the pricing is better. I'd still say it's a poor value if you need to buy an LCD and keyboard for it, though -- the price of all three items together gets up into iMac territory.
post #50 of 947
Quote:
Originally posted by melgross
There would be plenty of disagreements on that one, believe me. But, it's the ONLY high end board available for us, unfortunately.

I'd like for to show me a benchmark in a Pro 3D app where a QuadroFX 4500 didn't win.
onlooker
Registered User

Join Date: Dec 2001
Location: parts unknown




http://www.apple.com/feedback/macpro.html
Reply
onlooker
Registered User

Join Date: Dec 2001
Location: parts unknown




http://www.apple.com/feedback/macpro.html
Reply
post #51 of 947
Quote:
Originally posted by mjteix
with 2 options for each model:
- $1499 2.40GHz Conroe ($316)
- $1999 2.67GHz Conroe ($530) + better GPU/biggerHD/more RAM

- $2299 2.67GHz Woodcrest ($700) + better GPU
- $2499 3.00GHz Woodcrest ($850)

- $2999 dual 2.67GHz Woodcrest ($1400)
- $3299 dual 3.00GHz Woodcrest ($1700)

This is fine except that it's doubtful that Apple would use a single socket Woodcrest system. A 1333 MHz FSB and 3+ GHz single socket system is by definition a Conroe Extreme Edition system.

So:

$1599 2.40 GHz Conroe with 1066 FSB
$1999 2.66 GHz Conroe with 1066 FSB
$2599 2x2.66 GHz Woodcrest with 1066 FSB, maybe 1333 FSB
$3299 2x3.00 GHz Woodcrest with 1066 FSB, maybe 1333 FSB.

The troubling aspect will be the likely existence of $1000+, 3+ GHz, 1333 FSB Conroe Extreme Editions. There will be Conroe EE systems at 3+ GHz selling for the same price as Apple's standard config which will probably be one speed grade lower. I suppose users could upgrade by themselves.

The other kicker is the huge difference between a Conroe system and a Woodcrest system. It's so huge Apple should just make a midrange Conroe desktop and a professional Woodcrest desktop. One will be a mainstream single socket, single FSB system using 2 channel DDR2. The other will be a 2 socket, 2 FSB system using 4 channel FB-DIMM, a totally different memory architecture.

Quote:
Even before Woodcrest, Apple could build a quad Mac with the Yonah-based Sossaman ... It's not a 64-bit system, but there are some applications that are not 64-bit but are "core/processor" aware that would greatly benefit from quad-core (even @ 32-bit) systems...

If Apple wants to make a Intel quad today, they can simply use Pentium 4 CPUs, not much different from the G5 though. They either wait for Woodcrest, or use AMD.

Quote:
Will they be bold enough to please everybody with lower cost headless Macs?

No. They are interested in low margin products.
post #52 of 947
Some obvious predictions:

iMac 17": 2.0GHz Conroe LV (same thing as Merom, just ships earlier)
iMac 20": 2.3GHz Conroe LV

Good Mac Pro: 2.4GHz Conroe
Better Mac Pro: 3.0GHz Conroe
Best Mac Pro: 2x 3.0GHz Woodcrest
post #53 of 947
Quote:
Originally posted by onlooker
Even with the elegant design of the cube there was little demand for it, and it was thusly discontinued almost overnight.

I think that there's another factor in the discontinuation of the cube. When the cube came out, its 450 or 500 MHz G4 was the fastest available, the same processor that was in the G4 towers at the time. The revision of the G4 (to 733 MHz?) was too hot to be in the cube. Had Apple continued to sell it, the cube would not have been a small desktop Apple, but a small, slower desktop Apple. Or perhaps I've forgotten the actual sequence of events. My cube still runs, however.
post #54 of 947
I'm not saying the cube wasn't great. I loved it once I saw it in person. I was absolutely amazed with it. It wasn't the same as seeing it in a picture online, or in the paper. I imagine the Mac Mini would probably have a similar effect on me. All I'm saying is Apple tried it. Didn't work for them. And they pretty much have the bases covered without it. So don't get your hopes up. They covered it's typical demographic in their other machines.

There is one BIG THING you have to remember. What these current machines are not is designed from the beginning knowing they would have only an intel processor. The next ones will. The current ones are all legacy computers expecting to use IBM PPC's, and a possible intel transition, and their designs reflect that in terms of what they offer graphics wise etc. etc. etc. Now Apple can start thinking in different terms from the beginning about graphics, and actually use things that other PC manufacturers are using. Apple has a history of re-inventing computing. They didn't use a new design on any of these computers yet (like they normally would) because they intend to blow our minds now. You can count on that.
onlooker
Registered User

Join Date: Dec 2001
Location: parts unknown




http://www.apple.com/feedback/macpro.html
Reply
onlooker
Registered User

Join Date: Dec 2001
Location: parts unknown




http://www.apple.com/feedback/macpro.html
Reply
post #55 of 947
eweek reports Merom will have 20% performance improvement over Yonah with the same energy usage as Yonah.

Conroe will have 40% performance improvement with 40% energy reduction in contrast to Yonah.

Woodcrest will have 80% performance improvement vs 35% energy reduction in contrast to the current Xeon.

Intel intends for a dramatic difference between the two.
post #56 of 947
Quote:
Originally posted by TenoBell
eweek reports Merom will have 20% performance improvement over Yonah with the same energy usage as Yonah.

Conroe will have 40% performance improvement with 40% energy reduction in contrast to Yonah.

Woodcrest will have 80% performance improvement vs 35% energy reduction in contrast to the current Xeon.

Intel intends for a dramatic difference between the two.


And they should. The Netburst architecture was really losing steam. Anandtech's reviews showed that a 2Ghz Yonah was pretty much beating a Pentium D 3Ghz in almost every bench.

80% over current Xeon's is where Woodcrest needs to be. Besting Yonah by %80 would be a far more difficult task.
He's a mod so he has a few extra vBulletin privileges. That doesn't mean he should stop posting or should start acting like Digital Jesus.
- SolipsismX
Reply
He's a mod so he has a few extra vBulletin privileges. That doesn't mean he should stop posting or should start acting like Digital Jesus.
- SolipsismX
Reply
post #57 of 947
Quote:
Originally posted by onlooker
There is one BIG THING you have to remember. What these current machines are not is designed from the beginning knowing they would have only an intel processor. The next ones will. The current ones are all legacy computers expecting to use IBM PPC's, and a possible intel transition, and their designs reflect that in terms of what they offer graphics wise etc. etc. etc. Now Apple can start thinking in different terms from the beginning about graphics, and actually use things that other PC manufacturers are using. Apple has a history of re-inventing computing. They didn't use a new design on any of these computers yet (like they normally would) because they intend to blow our minds now. You can count on that. [/B]

This is very true, and people would do well to remember it.

These first machines are designed to make CURRENT Apple users comfortable with the idea of the Intel chip.

The next series, I hear, should be very different.
post #58 of 947
Quote:
Originally posted by hmurchison
And they should. The Netburst architecture was really losing steam. Anandtech's reviews showed that a 2Ghz Yonah was pretty much beating a Pentium D 3Ghz in almost every bench.

80% over current Xeon's is where Woodcrest needs to be. Besting Yonah by %80 would be a far more difficult task.

When they bring Hyperthreading back, things will get interesting.
post #59 of 947
Quote:
Originally posted by melgross
When they bring Hyperthreading back, things will get interesting.

I'm not sure they even need it. With the new architecture redesign they can probably ramp up a new faster processor faster than it would take them to try to restructure processors with hyper-threading. They came up with a whole new redesign for these processors because Moore's law had caught up with them. Moore's law is also why they were using hyper-threading. I always thought it was their last ditch effort to get more performance out of their designs as the speeds were not doubling any longer, but the number of the transistors were.
It may have also caused some of the problems. Only intel, and their R&D department know that for sure though; and that's if they actually tried to deduce what was hampering them in the first place.
Hyper-threading seemed to be their wanna be answer to altivec, but I always thought that the altivec SIMD was more effective for Mac users because of our use of photoshop, and that the majority of our developers were were smart enough to actually take advantage of the designs that were developed for use by the AIM.

Has intel even mentioned Hyper-Threading in their future plans?
onlooker
Registered User

Join Date: Dec 2001
Location: parts unknown




http://www.apple.com/feedback/macpro.html
Reply
onlooker
Registered User

Join Date: Dec 2001
Location: parts unknown




http://www.apple.com/feedback/macpro.html
Reply
post #60 of 947
Quote:
Originally posted by onlooker
I'm not sure they even need it. With the new architecture redesign they can probably ramp up a new faster processor faster than it would take them to try to restructure processors with hyper-threading. They came up with a whole new redesign for these processors because Moore's law had caught up with them. Moore's law is also why they were using hyper-threading. I always thought it was their last ditch effort to get more performance out of their designs as the speeds were not doubling any longer, but the number of the transistors were.
It may have also caused some of the problems. Only intel, and their R&D department know that for sure though; and that's if they actually tried to deduce what was hampering them in the first place.
Hyper-threading seemed to be their wanna be answer to altivec, but I always thought that the altivec SIMD was more effective for Mac users because of our use of photoshop, and that the majority of our developers were were smart enough to actually take advantage of the designs that were developed for use by the AIM.

Has intel even mentioned Hyper-Threading in their future plans?

I mentioned it because, they did.

Sometime in late 2007.
post #61 of 947
Seems inevitable Intel will need to bring back some form of multithreading.

Seems it would be more efficient to give two cores the ability to run like four or four the ability to run like eight, than to have actually have more cores.
post #62 of 947
If a cpu is able to feed all its working units properly without HT then you don't need it.
P4 has a problem with bubbles in it's extrem long pipeline. HT helped to fill them.
post #63 of 947
They pretty much have to have a quad woodcrest or continue production of the quad G5, perhaps even giving us a quad G5 upgrade. There is no way Apple will release "slower" machines than what they have now.

My pet theory is that Apple has very carefully managed clock-speed increases for the last few years. How unlikely is it that powerbooks go nicely from 1.67 GHz. to 1.8 GHz except on a new chip from a different company? It's suspiciously smooth and incremental. But then again, perhaps that's just the industry.

I predict a mix of Woodcrest and conroe in the PM with the conroe eventually falling into the iMac, the Kentsfield eventually taking over the high end. THere will never be an insanely spec'd workstation Mac because of clusted xServes. Hopefully, and most likely, the actual case of the next desktop will be smaller with more capacity.
post #64 of 947
"My pet theory is that Apple has very carefully managed clock-speed increases for the last few years."

And that promise to be at 3 GHz with the G5 in the summer of 2004 was a smoke screen

I try to not say anything about the G(5-1)
post #65 of 947
Thread Starter 
Quote:
Originally posted by wmf
Some obvious predictions:

iMac 17": 2.0GHz Conroe LV (same thing as Merom, just ships earlier)
iMac 20": 2.3GHz Conroe LV

Good Mac Pro: 2.4GHz Conroe
Better Mac Pro: 3.0GHz Conroe
Best Mac Pro: 2x 3.0GHz Woodcrest

I hope that's the way the Mac Pros work out because the single-3.0GHz is perfect for me, and I'll be naturally deterred without feeling like I'm buying inferior hardware if I don't get the dual 3.
post #66 of 947
Quote:
Originally posted by onlooker
There is one BIG THING you have to remember. What these current machines are not is designed from the beginning knowing they would have only an intel processor. The next ones will. The current ones are all legacy computers expecting to use IBM PPC's, and a possible intel transition, and their designs reflect that in terms of what they offer graphics wise etc. etc. etc.

I disagree on two counts. First, at the motherboard level these machines are all Intel because there is no commonality between Intel and PPC chips... Apple went entirely with the Intel CPUs and chipsets. The internal layouts of the iMac and MacBook Pro were substantially reorganized to accomodate the new designs.

Second, at the system design level the Intel chips aren't so remarkably more advanced that they enable all sorts of wonderful new applications. Yes, they deliver better performance than the G4 (although Apple never used the latest G4 so they could have gone a little farther there), and they can operate at lower power than the G5 (although Apple stopped pursuing IBM for better G5s). But the difference isn't so remarkable as to enable some wonderful new technology. And the main difference in terms of graphics is the integrated chipset graphics that everyone is complaining about.

It bears repeating that Apple switched to Intel for their long term roadmap, and I don't expect to see anything remarkable that they couldn't have done with PPC until well into 2007 or 2008. And remember, its easy to point at the last PPC machines they built and say "wow, look how much faster we are now" but then you are ignoring that the PPC would have improved had Apple stayed with it. No it wouldn't have improved by leaps and bounds (i.e. differences that have real impact), but then again neither have the Intel chips done that.

And for the comment that it took IBM 8 months to increase 200 MHz, you are overlooking the fact that Intel and AMD hardly progressed in terms of clock rate in the same period. In fact, the only really remarkable leap in the last couple of years (albeit somewhat specialized) is the Cell processor from you-know-who.

Short term the best thing about Apple using Intel is the large number of core variations and chipsets that they are producing, allowing more flexibility for Apple's various product lines. Long-term Intel's process technology looks good and will hopefully bring some truly remarkable products.
Providing grist for the rumour mill since 2001.
Reply
Providing grist for the rumour mill since 2001.
Reply
post #67 of 947
Quote:
Originally posted by smalM
If a cpu is able to feed all its working units properly without HT then you don't need it.
P4 has a problem with bubbles in it's extrem long pipeline. HT helped to fill them.

No. Simple as that.

They removed it temporarally for the new designs. But they will bring it back.

Intel can't keep adding cores to the chip.The expense and heat produced will, again, become unreasonable.

Read up on HT before you comment on it.
post #68 of 947
Quote:
Originally posted by Programmer
I disagree on two counts. First, at the motherboard level these machines are all Intel because there is no commonality between Intel and PPC chips... Apple went entirely with the Intel CPUs and chipsets. The internal layouts of the iMac and MacBook Pro were substantially reorganized to accomodate the new designs.

Second, at the system design level the Intel chips aren't so remarkably more advanced that they enable all sorts of wonderful new applications. Yes, they deliver better performance than the G4 (although Apple never used the latest G4 so they could have gone a little farther there), and they can operate at lower power than the G5 (although Apple stopped pursuing IBM for better G5s). But the difference isn't so remarkable as to enable some wonderful new technology. And the main difference in terms of graphics is the integrated chipset graphics that everyone is complaining about.

It bears repeating that Apple switched to Intel for their long term roadmap, and I don't expect to see anything remarkable that they couldn't have done with PPC until well into 2007 or 2008. And remember, its easy to point at the last PPC machines they built and say "wow, look how much faster we are now" but then you are ignoring that the PPC would have improved had Apple stayed with it. No it wouldn't have improved by leaps and bounds (i.e. differences that have real impact), but then again neither have the Intel chips done that.

And for the comment that it took IBM 8 months to increase 200 MHz, you are overlooking the fact that Intel and AMD hardly progressed in terms of clock rate in the same period. In fact, the only really remarkable leap in the last couple of years (albeit somewhat specialized) is the Cell processor from you-know-who.

Short term the best thing about Apple using Intel is the large number of core variations and chipsets that they are producing, allowing more flexibility for Apple's various product lines. Long-term Intel's process technology looks good and will hopefully bring some truly remarkable products.

I mostly agree with this.

But, I would emphasize the concept that both the G4 and 5 were at an impasse. I don't believe that Apple wasn't pushing to get more advanced chips. I feel that they were. But, as has been discussed ad nauseam, Both Freescale and IBM weren't interested. Apple's portion of their business wasn't increasing nearly as fast as the embedded portions were, and, of course, IBM had gotten several deals for other variants that looked to be much more profitable.

Freescale's dual core 600 series was far off, with a different memory model than they were using, and seemed to be too much trouble. That has also been discussed by Hannibal, and others.

The 7448 offered, at best, a marginal improvement that was simply not in the class of a dual Yonah, much less a dual Merom. so for portable use, Apple had no choice. Even had they gone to the 7448, they would have fallen even further behind this year.

The G5's could have held out for a while longer, but unless IBM has something up their sleeve that we don't know anything about, by the end of this year, they would have been severely outclassed. They are already outclassed when compared to the top Xeons, and noticeably so, by the equivalent Opterons. In video editing, Apple's strong point, the Opterons can do as much as a 30 to 40% faster render, under certain conditions. That HAD to be rectified.

IBM's refusal to even add to the cache, until late last year, even though the poor performance of the chips regarding latency has been criticized over and over again, shows that IBM has other markets in mind. Even a 1MB cache isn't really enough.
post #69 of 947
Quote:
Originally posted by Programmer
I disagree on two counts. First, at the motherboard level these machines are all Intel because there is no commonality between Intel and PPC chips... Apple went entirely with the Intel CPUs and chipsets. The internal layouts of the iMac and MacBook Pro were substantially reorganized to accomodate the new designs.

Second, at the system design level the Intel chips aren't so remarkably more advanced that they enable all sorts of wonderful new applications. Yes, they deliver better performance than the G4 (although Apple never used the latest G4 so they could have gone a little farther there), and they can operate at lower power than the G5 (although Apple stopped pursuing IBM for better G5s). But the difference isn't so remarkable as to enable some wonderful new technology. And the main difference in terms of graphics is the integrated chipset graphics that everyone is complaining about.

It bears repeating that Apple switched to Intel for their long term roadmap, and I don't expect to see anything remarkable that they couldn't have done with PPC until well into 2007 or 2008. And remember, its easy to point at the last PPC machines they built and say "wow, look how much faster we are now" but then you are ignoring that the PPC would have improved had Apple stayed with it. No it wouldn't have improved by leaps and bounds (i.e. differences that have real impact), but then again neither have the Intel chips done that.


After this point I am assuming you were addressing someone else because I didn't mention any of this
And for the comment that it took IBM 8 months to increase 200 MHz, you are overlooking the fact that Intel and AMD hardly progressed in terms of clock rate in the same period. In fact, the only really remarkable leap in the last couple of years (albeit somewhat specialized) is the Cell processor from you-know-who.

Short term the best thing about Apple using Intel is the large number of core variations and chipsets that they are producing, allowing more flexibility for Apple's various product lines. Long-term Intel's process technology looks good and will hopefully bring some truly remarkable products.

On points one, and two: My meaning was exactly about what you said. You just didn't get it. "The internal layouts of the iMac and MacBook Pro were substantially reorganized to accommodate the new designs. Yes That's exactly what my meaning was. Because they had to restrain the design to a previous make, and model size there were limits to what they could have done compared to designing a machine from the ground up. Integrated graphics may very well be in there because reorganizing a pre-existing laptop would not allow enough room to use laptop graphics cards. That is just one example. There is a rumored 17" version coming. This one is probably going to maintain the same design as the G4 version as well. In doing this Apple will gather a few switchers here and there, and I suspect we'll keep hearing complaints about the use of integrated graphics from users who would not switch because of it. I also suspect that when Apple decides to provide a new laptop design to us we will have laptop graphics cards in the pro models which will then give the allure of options to PC users, and Mac users who are not pleased with being stuck with an integrated graphics chipset. I am under the assumption that they will do the same in the iMac. I suspect that the Mac Mini, and maybe lower model iMacs will retain integrated graphics, but the MBP, 17" iMac will have a card slot. Apple new these transition Machines were just their launch pads, and they intend to make the most of the intel transition. They have proven that by starting early in releasing a non integrated version of boot-Camp outside of the "code named" Panther OS.
onlooker
Registered User

Join Date: Dec 2001
Location: parts unknown




http://www.apple.com/feedback/macpro.html
Reply
onlooker
Registered User

Join Date: Dec 2001
Location: parts unknown




http://www.apple.com/feedback/macpro.html
Reply
post #70 of 947
Quote:
Originally posted by melgross
Read up on HT before you comment on it.

Thanks Massa for the advice to us stupids!
post #71 of 947
Quote:
Originally posted by onlooker
Because they had to restrain the design to a previous make, and model size there were limits to what they could have done compared to designing a machine from the ground up. Integrated graphics may very well be in there because reorganizing a pre-existing laptop would not allow enough room to use laptop graphics cards. That is just one example. There is a rumored 17" version coming. This one is probably going to maintain the same design as the G4 version as well. In doing this Apple will gather a few switchers here and there, and I suspect we'll keep hearing complaints about the use of integrated graphics from users who would not switch because of it. I also suspect that when Apple decides to provide a new laptop design to us we will have laptop graphics cards in the pro models which will then give the allure of options to PC users, and Mac users who are not pleased with being stuck with an integrated graphics chipset. I am under the assumption that they will do the same in the iMac. I suspect that the Mac Mini, and maybe lower model iMacs will retain integrated graphics, but the MBP, 17" iMac will have a card slot. Apple new these transition Machines were just their launch pads, and they intend to make the most of the intel transition. They have proven that by starting early in releasing a non integrated version of boot-Camp outside of the "code named" Panther OS. [/B]

I'm not sure of what you're saying here.

When we talk "integrated graphics" we mean Intel's chip.

This is clearly not what is being used in the MBP or the iMac, both of which use ATI 1600's. That is a far better graphics solution.

But, I see that you are talking about a slot. You don't think that Apple will make the graphics subsystem in the MBP and iMac to be upgradeable, do you? Because that's very unlikely to happen.

What PC laptops have removeable graphics cards, other than, possibly, the very expensive, and heavy, machines that a few small specialty companies sell?
post #72 of 947
Quote:
Originally posted by smalM
Thanks Massa for the advice to us stupids!

Always glad to be of service.

But, you called yourself stupid. I didn't. Uninformed, yes.

The advice still stands.
post #73 of 947
Quote:
Originally posted by DrBoar
"My pet theory is that Apple has very carefully managed clock-speed increases for the last few years."

And that promise to be at 3 GHz with the G5 in the summer of 2004 was a smoke screen

I try to not say anything about the G(5-1)

I guess I should have said "couple of years". I think IBM told Apple they would hit 3 GHz easily and there would be a lot more good stuff after that. When IBM missed and, presumably, scaled back their roadmap (putting their efforts into the cell processor) that's when Apple decided to go with Intel. I'm sure Apple has always been in communication with Intel and kind of knew what was coming.

Since Apple was the only G5 customer and the only major G4 customer they could kind of dictate what came out when. So they planned this gradual transition. makes good sence to do it that way.
post #74 of 947
I see it as, Well Apple here is your Dual 2GHZ chip, great deal huh, we're good eh; now here's your 2.5GHz chip but looks like you'll have to water-cool it...and your 3 GHz chip is a piece of cake, oh, did we mention you'll have to cool it with liquid Helium?...but don't worry we gonna make these nice Cell chips for ya, but MS and others are in the que in front of ya to get them...you don't mind waiting do you...

Steve looks over both shoulders, looks at the IBM suits and tells them: kiss my a$$...
I heard that geeks are a dime a dozen, I just want to find out who's been passin' out the dimes
----- Fred Blassie 1964
Reply
I heard that geeks are a dime a dozen, I just want to find out who's been passin' out the dimes
----- Fred Blassie 1964
Reply
post #75 of 947
I think hardware multithreading is pretty much relegated to a niche server market, the only place where it could be of real value, and you won't see it on personal computers again anytime soon.

This year, 2006, there are dual-core CPUs aplenty. In 2007, quad-cores start shipping and in 2008 there will be quite a few of them coming out of 45 nm fabs. In 2010, octo-cores can come out if 32 nm pans out.

Why have multithreading when so many cores could be available in 2 to 4 years? On a high-end 2 socket system, that could be 8 to 16 cores. What sort of software needs that many threads? (Other than server related stuff.)

So, there are few of things holding multithreading back on personal computers: multi-cores are and will be available, the system architecture will be I/O bound for multi-cores let alone MT multi-cores, and software development cycles seem to operate over longer durations that fab development cycles.

MT is nice for servers where I/O bandwidth is more plentiful (and more expensive), where lots of small chunks of CPU time are required to be delivered. Intel may bring MT back in Woodcrest, but I don't see anything driving it in personal computers. It may even be multi-cores (CMP) drive MT out of server processors.

As for Cell and Xenon/Waternoose, I see that as a freebie courtesy of IBM, or something that they hope pans out 3 to 4 years into the future. Ie, the mission of game consoles is to sell games. New games can gradually tap the resources of the hardware translating to more games being sold, and a game console sold at a big loss is no problem as users continue to buy 3 or 4 games a year (or pay subscription fees). So, MT eventually will be useful.
post #76 of 947
Quote:
Originally posted by DrBoar
"My pet theory is that Apple has very carefully managed clock-speed increases for the last few years."

And that promise to be at 3 GHz with the G5 in the summer of 2004 was a smoke screen

I try to not say anything about the G(5-1)

Intel actually delivers without the need for watercooling.
"Humankind -- despite its artistic pretensions, its sophistication, and its many accomplishments -- owes its existence to a six-inch layer of topsoil and the fact that it rains."
Reply
"Humankind -- despite its artistic pretensions, its sophistication, and its many accomplishments -- owes its existence to a six-inch layer of topsoil and the fact that it rains."
Reply
post #77 of 947
Quote:
Originally posted by theapplegenius
Intel actually delivers without the need for watercooling.

It's amazing at just how many watercooling upgrades are out there though.

Also, it was stated a while ago that Intel was investigating watercooling as well.
post #78 of 947
Quote:
Originally posted by melgross
I'm not sure of what you're saying here.

When we talk "integrated graphics" we mean Intel's chip.

This is clearly not what is being used in the MBP or the iMac, both of which use ATI 1600's. That is a far better graphics solution.

But, I see that you are talking about a slot. You don't think that Apple will make the graphics subsystem in the MBP and iMac to be upgradeable, do you? Because that's very unlikely to happen.

What PC laptops have removeable graphics cards, other than, possibly, the very expensive, and heavy, machines that a few small specialty companies sell?

My bad. Crossed my thoughts with the mini's integrated graphics chip, but your right that I was referring to an upgradable graphics set up for both systems. So on to it. The MBP is less attractive than other performance laptops in the same range (others are cheaper) like the Alienware Area-51® m7700, and now that DELL owns Alienware they are going to market every thing as "Better than Apple" in as "many ways" as they can. For pro's looking to use laptops most would prefer to have options, and these are huge options. Options that can make a big difference that separate working situations like 2D from 3D. Gamers like to have the same options, and with windows running on the Mac both windows and Mac users/gamers having easy access to windows games on Macs things need to be shaken up, and options can make all the difference.
If Apple doesn't open up to upgradable graphics choice is what will keep the PC separate from the Mac experience, and keep the PC camp on the PC side.
Word of mouth marketing will remind PC users that on a PC they have options of going from a GeForce Go 7300, to a GeForce Go 7800 GTX, or for 3D the NVIDIA Quadro FX 2500M; (and this doesn't even include ATI cards) And that's not all. All these cards can be used in laptops using SLI. Nvidia has partnered with 20 Laptop makers to offer SLI graphics in notebooks.
The best choice for editing HD video content on a Laptop would be a SLI enabled laptop with a pair of NVIDIA GeForce Go 7800 GTX GPUs using the NVIDIA nForce4 SLI media communications processor (MCP). Things like that will keep pro video editors using laptops with Adobe Premiere Pro, and After effects instead of FCP-HD, After effects, and Motion.

On to the iMac. (I'll keep it short)
People keep asking for the (failed) headless Mac again because the iMac graphics isn't enough, and it's not upgradable. If Apple offered upgradable graphics cards in the iMac I'd say 70% of them would forget about it. As for size they could easily use Go series cards for laptops, and outperform the current ones.

It's obvious that Apple is going directly after the biggest piece of the PC market that they can get their hands on. If they want to get half of what they would like of it they need to show PC users that Apple know's what they want, and Apple can provide it. Hell half of this shit Mac users want too. Believe me. I would love to have the option of SLI in a PowerMac, but that's another thing entirely.
onlooker
Registered User

Join Date: Dec 2001
Location: parts unknown




http://www.apple.com/feedback/macpro.html
Reply
onlooker
Registered User

Join Date: Dec 2001
Location: parts unknown




http://www.apple.com/feedback/macpro.html
Reply
post #79 of 947
Quote:
Originally posted by onlooker
My bad. Crossed my thoughts with the mini's integrated graphics chip, but your right that I was referring to an upgradable graphics set up for both systems. So on to it. The MBP is less attractive than other performance laptops in the same range (others are cheaper) like the Alienware Area-51® m7700, and now that DELL owns Alienware they are going to market every thing as "Better than Apple" in as "many ways" as they can. For pro's looking to use laptops most would prefer to have options, and these are huge options. Options that can make a big difference that separate working situations like 2D from 3D. Gamers like to have the same options, and with windows running on the Mac both windows and Mac users/gamers having easy access to windows games on Macs things need to be shaken up, and options can make all the difference.
If Apple doesn't open up to upgradable graphics choice is what will keep the PC separate from the Mac experience, and keep the PC camp on the PC side.
Word of mouth marketing will remind PC users that on a PC they have options of going from a GeForce Go 7300, to a GeForce Go 7800 GTX, or for 3D the NVIDIA Quadro FX 2500M; (and this doesn't even include ATI cards) And that's not all. All these cards can be used in laptops using SLI. Nvidia has partnered with 20 Laptop makers to offer SLI graphics in notebooks.
The best choice for editing HD video content on a Laptop would be a SLI enabled laptop with a pair of NVIDIA GeForce Go 7800 GTX GPUs using the NVIDIA nForce4 SLI media communications processor (MCP). Things like that will keep pro video editors using laptops with Adobe Premiere Pro, and After effects instead of FCP-HD, After effects, and Motion.

On to the iMac. (I'll keep it short)
People keep asking for the (failed) headless Mac again because the iMac graphics isn't enough, and it's not upgradable. If Apple offered upgradable graphics cards in the iMac I'd say 70% of them would forget about it. As for size they could easily use Go series cards for laptops, and outperform the current ones.

It's obvious that Apple is going directly after the biggest piece of the PC market that they can get their hands on. If they want to get half of what they would like of it they need to show PC users that Apple know's what they want, and Apple can provide it. Hell half of this shit Mac users want too. Believe me. I would love to have the option of SLI in a PowerMac, but that's another thing entirely.

You're speaking about build to order options from these companies? Not user replaceable boards.

The largest part of the PC market doesn't change their graphics cards. Whatever comes with the machine, dies with the machine,.
post #80 of 947
Quote:
Originally posted by melgross
It's amazing at just how many watercooling upgrades are out there though.

Also, it was stated a while ago that Intel was investigating watercooling as well.

Watercooling is great, I have it in my PC. But it shouldn't be necessary.
"Humankind -- despite its artistic pretensions, its sophistication, and its many accomplishments -- owes its existence to a six-inch layer of topsoil and the fact that it rains."
Reply
"Humankind -- despite its artistic pretensions, its sophistication, and its many accomplishments -- owes its existence to a six-inch layer of topsoil and the fact that it rains."
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Future Apple Hardware
AppleInsider › Forums › Mac Hardware › Future Apple Hardware › The Intel Powermac / Powermac Conroe / Mac Pro thread